Campaign contributions and voting on the cargo preference bill: A comparison of simultaneous models

Abstract

A variety of single equation and simultaneous equation models have been employed to test the hypothesis that interest group campaign contributions influence congressional voting. Although contributions exhibited the anticipated positive effect in all models, the simultaneous equations estimates generally indicated much lower significance levels for the contribution coefficients than did the single equation models. The lower significance levels are apparently attributable to a lack of precision of the simultaneous model estimates (indicated by large standard errors) as well as possible bias of the single equation models. It was also found that qualitative results from the more sophisticated simultaneous probit-Tobit models were quite similar to those obtained from 2SLS estimation of the linear probability model.